International Conference on Renewable Energies and Power Quality (ICREPQ’16)
Madrid (Spain), 4th to 6th May, 2016 Renewable Energy and Power Quality Journal(RE&PQJ)
ISSN 2172-038 X, No.14 May 2016
Optical high voltage breakdown prediction
using thermal lensing effect in transformer oil
R. Struebig and I. Glesk
Department of Electronic and Electrical Engineering
University of Strathclyde
204 George street, Glasgow G11XW, Scotland, UK
Phone/Fax number: +441415482529/+44-141-552-4968, e-mail: [email protected], [email protected]
Abstract. We describe an optical system to monitor
microscopic pre-breakdown events in liquid insulation. The
system has successfully demonstrated its ability to predict high
voltage breakdown in transformer oil. A simple theory based on
a thermal lens build up between electrodes as a result of applied
voltage is presented to explain the system operation and obtained
measured results.
Key words
Breakdown prediction, Liquid insulator, High voltage, Oil
movement, Laser measurements.
1. Introduction
Expensive assets such as power transformers are vital
parts of modern power transmission networks. The value
of network components lies not only in their price but also
in the loss of profit if they become in operational due to
the damage.
Therefore, it is crucial to develop new and effective
methods for their condition monitoring and the ability to
predict high voltage (HV) breakdowns. This will allow
network providers to improve the reliability and reduce
breakdown costs through appropriate maintenance [1].
The aim of this paper is to demonstrate a simple optical
system for monitoring and electrical breakdown prediction
in liquid insulation.
2. Pre breakdown events in liquid insulation
Stress induced oil movements by a strong electric field E
where recorded in earlier research [2]. The conclusion
drawn was that the oil movement is induced due to a
moving space charge layer injected at the HV
electrodes [3].
Takashima [4] using computer modelling determined the
oil velocity in an oil filled vessel and developed the theory
of a cylindrically shaped charge channel (see Fig. 1). This
was later also observed by Atten [5]. The channel with
radius αc is extending from the tip of the electrode 1 to the
plane electrode 2 and is created by the applied electric
field E. Takashima has shown that αc as well as the oil
speed are constant over nearly the full length of the charge
channel. His findings draw a correlation between the
applied HV, the oil velocity, and the radius of the oil
channel.
Figure 1: Charge channel generated in electrically stressed
transformer oil. [4]
A bigger needle tip will also increase αc while reducing
the speed of oil movement in the charge channel [4].
https://doi.org/10.24084/repqj14.295 290 RE&PQJ, No.14, May 2016
As the result of the applied HV a current flow was
observed this leads to energy dissipation into the oil and
some of this energy will be transformed into heat [4, 5]
subsequently warming the oil.
Such electric stress will lead to local breakdown in oil.
This is known as partial discharge, PD [6]. A presence of
the electric current due to discharge events is another
factor leading to the localized oil heating [6].
By summarizing the above one can conclude the
temperature distribution in oil can be represented by a
thermal field resulted from events triggered by HV
applied to electrodes.
We will show the effect of HV on the transformer oil
temperature variations and changes in the oils index of
refraction. By passing a laser beam between the electrodes
through the charge channel. Analysing the changes of the
laser beam properties we can visualize microscopic pre
breakdown effects on the passing laser beam and predict
an ‘approaching’ HV breakdown in oil.
3. Experimental investigation of breakdown
in oil transformer
In order to exploit these events for HV breakdown
prediction in transformer oil we have built an
experimental setup shown in Fig. 2. The setup consists of
a transparent cylindrical vessel (Fig. 2a) filled with
transformer oil. Two electrodes (needle and flat) are fitted
through the vessels walls. The gap between electrodes was
set to 1.5 cm. The needle tip radius was 1.5 mm.
Figure 2: Measurement system setup a) top view of the setup; b)
Screen to observe laser beam
A laser pointer was placed outside the vessel. The laser
pointer was positioned 12.5 cm away from the midpoint of
the oil vessel and adjusted to pass the laser beam through
the oil between the vessels electrodes before hitting the
screen (Fig. 2b). The total beam path length was 7.9 m
(measured from the electrodes center point to the screen).
In each experiment the high voltage applied to the
electrodes was raised in steps from 0V to 8640V, 12880V,
17520V, 22200V, 26680V, and 31470V and kept constant
for a minimum of 40 seconds.
The voltage was recorded in real time by a digital
oscilloscope using the BenchVue software package. Beam
shape changes on the screen were recorded by a camera.
From the recorded data we plotted changes of the laser
beam width on the screen with any change of HV applied
to electrodes. To record these beam width changes the
Kinovea software package was used. The Software
recorded the changing beam width shown in Fig. 4 by
tracking the laser beam borders movement as shown in
Fig. 3 as a function of time parallel to the applied HV
across electrodes. The red (green) line respectively,
represents the upper (lower) laser beam border the
position is measured in pixels (px, 1 px = 0.263 mm) and
is recorded on the lefty axis in Fig. 3.
Figure 3: Change in the upper (red) and lover (green) laser beam
borders position on the screen due to chances in applied voltage
(blue)
Figure 4: Changes in Laser beam width (red) due to chances in
applied voltage (blue)
https://doi.org/10.24084/repqj14.295 291 RE&PQJ, No.14, May 2016
The corresponding applied voltage in kilovolt (kV) is the
blue trace measured on the right y axis.
From the data analyses in Fig. 3 it can be said that both
laser borders are stable until 12.9kV. For higher voltage
values both borders start to change positions but it is
mainly the lower laser border that moves considerably as
the HV increases which is a result of experimental
arrangements. After the fifth voltage increase the lower
border also stops moving but instead starts to oscillate and
the insulation breakdown follows as it is clearly visible in
Fig. 4. This behaviour was consistent in all our repeated
experiments.
4. Discussion
The observed effects could be explained by taking into
account the thermal stress in oil that is resulted from HV
applied across the electrodes. The stress creates a
dynamically changing cylindrical lens alongside the
charge channel (see Fig. 5) and causes a thermal lensing
effect in oil which in turn affects the passing laser beam.
We applied a geometrical approach leading to a lens
maker’s formula for a thin lens [7].
1
𝑓(𝑉)= (
𝑛𝑙𝑒𝑛𝑠(𝑉) − 𝑛𝑜𝑖𝑙𝑛𝑜𝑖𝑙
)2
𝑅(𝑉) (1)
Where f(V) is the focal length of this thermal cylindrical
lens, R(V) its radii of curvature and nlens(V) is the index of
refraction represented by the refractive index of the
warming up oil in the charge channel. It must be noted
that all these three parameter are functions of the applied
voltage V. Were by noil is the refractive index of the oil
outside of the charge channel (noil = 1.47583 for 20°C [8])
and its value is constant. Any changes of the lens
parameters via voltage V applied to the electrodes are
directly responsible for shifts of the laser beam boundaries
recorded and seen in Fig. 3 and Fig. 4. It can be easily
shown that a change of the focal length f will result in the
change of the laser beam image width hi observed on the
screen (see Fig. 2).
The ‘creation’ of this thermal lens is a direct result of
changing properties of the oils index inside the charge
cylinder by the applied voltage. The oil movement
observed by Takashima and Atten [4, 5] is a contributing
factor codifying the lens diameter. In first approximation
the charge channel diameter αc could be viewed identical
with the lens parameter R. According to Takashima’s
observation the charge channel diameter αc can increase
up to 6 times with the increasing voltage [4]. Atten also
observed the increase in the channel diameter with voltage
[5]. This directly supports our proposal that increases in
the charge channel diameter directly influence the lens
radius R, thus its f via Eq. 1, and are results of changing
HV.
We know that the refractive index of oil decreases with
the temperature (-3.75×10-4dn/dT °C-1) [8]. Any increases
in HV lead to higher currents [3] in the charge channel
and will cause oil heating in the channel thus leading to
voltage dependent changes of the refractive index inside
the channel. This heat field takes the shape of the moving
oil and is directly responsible for creation of the described
cylindrically shaped lens like an object alongside the
channel. Stishkov simulated and afterwards observed the
geometry of the moving oil [9] and came to similar results
as Takashima. The lens refraction index
nlens(V) = nhot-oil(V) closely follows oil temperature
changes within the charge channel, from the hottest value
Figure 5: Laser beam affected by a thermal gradient lens created
by a charge channel in HV stressed insulation between
electrodes
(lowest index of refraction) in its middle to lower
temperatures (higher indexes) in outward directions
according to the oil speed contours, until the temperature
equilibrium with the oil surrounding the charge channel is
reached. Because these are HV related changes of the lens
parameters they are directly responsible for the observed
beam movements seen on the screen and based on our
measurements quantified in Fig. 3 and Fig. 4.
When it comes to HV breakdown prediction, just before
this event, partial discharges of high energy lead to the
rapid development of gas channels between the electrodes
[6]. These create shockwaves [6] and lead to the
disruption of the laminar flow between the electrodes
[4, 9] within the charge channel (the host of the discussed
thermal cylindrical lens). We expect that these events are
disrupting the uniformity of this lens strongly affecting the
passing laser beam. This manifest itself as violent laser
beam movements observed on the screen just before the
breakdown point as recorded in Fig. 4. The movement is
mainly due to the change in the lens geometry and local
changes in nlens due to the additional heating from the PD
event. The lens then recovers to its initial geometry due to
the moving oil closing the gas channels and carrying away
or dissolving the additional heat. By observing the laser
beam on the screen we were able to predict the point of
total insulation breakdown.
5. Conclusion
We described a measurement system which uses a simple
optical concept for monitoring microscopic pre-
breakdown events in liquid insulators such as oil. The
system has been built and demonstrated in the laboratory
and is capable of predicting high voltage breakdowns. A
simple theory based on a thermal lens build up between
electrodes within the charge channel as a result of HV is
presented to explain the system operation and observed
measurement results.
https://doi.org/10.24084/repqj14.295 292 RE&PQJ, No.14, May 2016
References
[1] M. Judd, L. Yang, I. Hunter, “Partial Discharge Monitoring
for Power Transformers Using UHF Sensors Part 1: Sensors and
Signal Interpretation,” IEEE Electrical Insulation Magazine,
Vol. 21, No.2; March/April 2005
[2] J. Cross, M. Rakano, S. Savannis, “Electric Stress Induced
motion in Transformer Oils under 60 Hz Stress, “Journal of
Electrostatics, 7 page 361--372 361; Elsevier Scientific
Publishing Company, Amsterdam; 1979
[3] P. Atten, M. Haidara, “Electrical Conduction and EHD
Motion of Dielectric Liquids in a Knife-plane electrode
Assembly,” IEEE Transactions on Electrical Insulation Vol. EI-
20 No.2, April 1965
[4] T. Takashima, “I-V Characteristics and Liquid Motion in
Needle-to-Plane and Razor Blade- to- Plane Configurations in
Transformer Oil and Liquid Nitrogen,” IEEE Transactions on
Electrical Insulation Vol. 23 No. 4, August 1988, page 645-658
[5] P. Atten, B. Malraison, M. Zahn, “Electro hydrodynamic
Plumes in Point- plane Geometry,” IEEE Transactions on
Dielectrics and Electrical Insulation, Vol. 4 No. 6, 1997
[6] M. Butcher at al, “Conduction and Breakdown Mechanisms
in Transformer Oil”; IEEE Transactions on plasma science, Vol.
3, No 2; April 2006
[7] J. Greivenkamp, “Field Guide to Geometrical Optics,” SPIE
Press Field Guides Vol FG01; Washington, USA; 2004
[8] L. Watters, “Liquid Refractive Index - Mineral Oil,”
Department of commerce USA, National Institute of Standards
&Technology, Certificate, Standard Reference Material® 1922
[9] Yu.K. Stishkov, V.A. Chirkov, “Features of electro
hydrodynamic flows in needle- plane electrode system,” IEEE
International Conference, Dielectric Liquids, 2008. ICDL 2008
https://doi.org/10.24084/repqj14.295 293 RE&PQJ, No.14, May 2016